]>
Commit | Line | Data |
---|---|---|
ef416fc2 | 1 | /* |
bc44d920 | 2 | * "$Id: search.c 6649 2007-07-11 21:46:42Z mike $" |
ef416fc2 | 3 | * |
4 | * Search routines for the Common UNIX Printing System (CUPS). | |
5 | * | |
91c84a35 | 6 | * Copyright 2007-2008 by Apple Inc. |
07725fee | 7 | * Copyright 1997-2006 by Easy Software Products. |
ef416fc2 | 8 | * |
9 | * These coded instructions, statements, and computer programs are the | |
bc44d920 | 10 | * property of Apple Inc. and are protected by Federal copyright |
11 | * law. Distribution and use rights are outlined in the file "LICENSE.txt" | |
12 | * which should have been included with this file. If this file is | |
13 | * file is missing or damaged, see the license at "http://www.cups.org/". | |
ef416fc2 | 14 | * |
15 | * Contents: | |
16 | * | |
17 | * cgiCompileSearch() - Compile a search string. | |
18 | * cgiDoSearch() - Do a search of some text. | |
19 | * cgiFreeSearch() - Free a compiled search context. | |
20 | */ | |
21 | ||
22 | /* | |
23 | * Include necessary headers... | |
24 | */ | |
25 | ||
26 | #include "cgi-private.h" | |
27 | #include <regex.h> | |
28 | ||
29 | ||
30 | /* | |
31 | * 'cgiCompileSearch()' - Compile a search string. | |
32 | */ | |
33 | ||
34 | void * /* O - Search context */ | |
35 | cgiCompileSearch(const char *query) /* I - Query string */ | |
36 | { | |
37 | regex_t *re; /* Regular expression */ | |
38 | char *s, /* Regular expression string */ | |
39 | *sptr, /* Pointer into RE string */ | |
40 | *sword; /* Pointer to start of word */ | |
41 | int slen; /* Allocated size of RE string */ | |
42 | const char *qptr, /* Pointer into query string */ | |
43 | *qend; /* End of current word */ | |
44 | const char *prefix; /* Prefix to add to next word */ | |
45 | int quoted; /* Word is quoted */ | |
46 | int wlen; /* Word length */ | |
47 | char *lword; /* Last word in query */ | |
48 | ||
49 | ||
50 | DEBUG_printf(("help_compile_search(query=\"%s\")\n", query ? query : "(nil)")); | |
51 | ||
52 | /* | |
53 | * Allocate a regular expression storage structure... | |
54 | */ | |
55 | ||
91c84a35 MS |
56 | if ((re = (regex_t *)calloc(1, sizeof(regex_t))) == NULL) |
57 | return (NULL); | |
ef416fc2 | 58 | |
59 | /* | |
60 | * Allocate a buffer to hold the regular expression string, starting | |
61 | * at 1024 bytes or 3 times the length of the query string, whichever | |
62 | * is greater. We'll expand the string as needed... | |
63 | */ | |
64 | ||
65 | slen = strlen(query) * 3; | |
66 | if (slen < 1024) | |
67 | slen = 1024; | |
68 | ||
91c84a35 MS |
69 | if ((s = (char *)malloc(slen)) == NULL) |
70 | { | |
71 | free(re); | |
72 | return (NULL); | |
73 | } | |
ef416fc2 | 74 | |
75 | /* | |
76 | * Copy the query string to the regular expression, handling basic | |
77 | * AND and OR logic... | |
78 | */ | |
79 | ||
80 | prefix = ".*"; | |
81 | qptr = query; | |
82 | sptr = s; | |
83 | lword = NULL; | |
84 | ||
85 | while (*qptr) | |
86 | { | |
87 | /* | |
88 | * Skip leading whitespace... | |
89 | */ | |
90 | ||
91 | while (isspace(*qptr & 255)) | |
92 | qptr ++; | |
93 | ||
94 | if (!*qptr) | |
95 | break; | |
96 | ||
97 | /* | |
98 | * Find the end of the current word... | |
99 | */ | |
100 | ||
101 | if (*qptr == '\"' || *qptr == '\'') | |
102 | { | |
103 | /* | |
104 | * Scan quoted string... | |
105 | */ | |
106 | ||
107 | quoted = *qptr ++; | |
108 | for (qend = qptr; *qend && *qend != quoted; qend ++); | |
109 | ||
110 | if (!*qend) | |
111 | { | |
112 | /* | |
113 | * No closing quote, error out! | |
114 | */ | |
115 | ||
116 | free(s); | |
117 | free(re); | |
118 | ||
119 | if (lword) | |
120 | free(lword); | |
121 | ||
122 | return (NULL); | |
123 | } | |
124 | } | |
125 | else | |
126 | { | |
127 | /* | |
128 | * Scan whitespace-delimited string... | |
129 | */ | |
130 | ||
131 | quoted = 0; | |
132 | for (qend = qptr + 1; *qend && !isspace(*qend); qend ++); | |
133 | } | |
134 | ||
135 | wlen = qend - qptr; | |
136 | ||
137 | /* | |
138 | * Look for logic words: AND, OR | |
139 | */ | |
140 | ||
141 | if (wlen == 3 && !strncasecmp(qptr, "AND", 3)) | |
142 | { | |
143 | /* | |
144 | * Logical AND with the following text... | |
145 | */ | |
146 | ||
147 | if (sptr > s) | |
148 | prefix = ".*"; | |
149 | ||
150 | qptr = qend; | |
151 | } | |
152 | else if (wlen == 2 && !strncasecmp(qptr, "OR", 2)) | |
153 | { | |
154 | /* | |
155 | * Logical OR with the following text... | |
156 | */ | |
157 | ||
158 | if (sptr > s) | |
159 | prefix = ".*|.*"; | |
160 | ||
161 | qptr = qend; | |
162 | } | |
163 | else | |
164 | { | |
165 | /* | |
166 | * Add a search word, making sure we have enough room for the | |
167 | * string + RE overhead... | |
168 | */ | |
169 | ||
170 | wlen = (sptr - s) + 4 * wlen + 2 * strlen(prefix) + 4; | |
171 | ||
172 | if (wlen > slen) | |
173 | { | |
174 | /* | |
175 | * Expand the RE string buffer... | |
176 | */ | |
177 | ||
178 | char *temp; /* Temporary string pointer */ | |
179 | ||
180 | ||
181 | slen = wlen + 128; | |
182 | temp = (char *)realloc(s, slen); | |
183 | if (!temp) | |
184 | { | |
185 | free(s); | |
186 | free(re); | |
187 | ||
188 | if (lword) | |
189 | free(lword); | |
190 | ||
191 | return (NULL); | |
192 | } | |
193 | ||
194 | sptr = temp + (sptr - s); | |
195 | s = temp; | |
196 | } | |
197 | ||
198 | /* | |
199 | * Add the prefix string... | |
200 | */ | |
201 | ||
202 | strcpy(sptr, prefix); | |
203 | sptr += strlen(sptr); | |
204 | ||
205 | /* | |
206 | * Then quote the remaining word characters as needed for the | |
207 | * RE... | |
208 | */ | |
209 | ||
210 | sword = sptr; | |
211 | ||
212 | while (qptr < qend) | |
213 | { | |
214 | /* | |
215 | * Quote: ^ . [ $ ( ) | * + ? { \ | |
216 | */ | |
217 | ||
218 | if (strchr("^.[$()|*+?{\\", *qptr)) | |
219 | *sptr++ = '\\'; | |
220 | ||
221 | *sptr++ = *qptr++; | |
222 | } | |
223 | ||
07725fee | 224 | *sptr = '\0'; |
225 | ||
ef416fc2 | 226 | /* |
227 | * For "word1 AND word2", add reciprocal "word2 AND word1"... | |
228 | */ | |
229 | ||
230 | if (!strcmp(prefix, ".*") && lword) | |
231 | { | |
232 | char *lword2; /* New "last word" */ | |
233 | ||
234 | ||
91c84a35 MS |
235 | if ((lword2 = strdup(sword)) == NULL) |
236 | { | |
237 | free(lword); | |
238 | free(s); | |
239 | free(re); | |
240 | return (NULL); | |
241 | } | |
ef416fc2 | 242 | |
243 | strcpy(sptr, ".*|.*"); | |
244 | sptr += 5; | |
245 | ||
246 | strcpy(sptr, lword2); | |
247 | sptr += strlen(sptr); | |
248 | ||
249 | strcpy(sptr, ".*"); | |
250 | sptr += 2; | |
251 | ||
252 | strcpy(sptr, lword); | |
253 | sptr += strlen(sptr); | |
254 | ||
255 | free(lword); | |
256 | lword = lword2; | |
257 | } | |
258 | else | |
259 | { | |
260 | if (lword) | |
261 | free(lword); | |
262 | ||
263 | lword = strdup(sword); | |
264 | } | |
265 | ||
266 | prefix = ".*|.*"; | |
267 | } | |
268 | ||
269 | /* | |
270 | * Advance to the next string... | |
271 | */ | |
272 | ||
273 | if (quoted) | |
274 | qptr ++; | |
275 | } | |
276 | ||
277 | if (lword) | |
278 | free(lword); | |
279 | ||
280 | if (sptr > s) | |
281 | strcpy(sptr, ".*"); | |
282 | else | |
283 | { | |
284 | /* | |
285 | * No query data, return NULL... | |
286 | */ | |
287 | ||
288 | free(s); | |
289 | free(re); | |
290 | ||
291 | return (NULL); | |
292 | } | |
293 | ||
294 | /* | |
295 | * Compile the regular expression... | |
296 | */ | |
297 | ||
298 | DEBUG_printf((" s=\"%s\"\n", s)); | |
299 | ||
300 | if (regcomp(re, s, REG_EXTENDED | REG_ICASE)) | |
301 | { | |
302 | free(re); | |
303 | free(s); | |
304 | ||
305 | return (NULL); | |
306 | } | |
307 | ||
308 | /* | |
309 | * Free the RE string and return the new regular expression we compiled... | |
310 | */ | |
311 | ||
312 | free(s); | |
313 | ||
314 | return ((void *)re); | |
315 | } | |
316 | ||
317 | ||
318 | /* | |
319 | * 'cgiDoSearch()' - Do a search of some text. | |
320 | */ | |
321 | ||
322 | int /* O - Number of matches */ | |
323 | cgiDoSearch(void *search, /* I - Search context */ | |
324 | const char *text) /* I - Text to search */ | |
325 | { | |
326 | int i; /* Looping var */ | |
327 | regmatch_t matches[100]; /* RE matches */ | |
328 | ||
329 | ||
330 | /* | |
331 | * Range check... | |
332 | */ | |
333 | ||
334 | if (!search || !text) | |
335 | return (0); | |
336 | ||
337 | /* | |
338 | * Do a lookup... | |
339 | */ | |
340 | ||
341 | if (!regexec((regex_t *)search, text, sizeof(matches) / sizeof(matches[0]), | |
342 | matches, 0)) | |
343 | { | |
344 | /* | |
345 | * Figure out the number of matches in the string... | |
346 | */ | |
347 | ||
348 | for (i = 0; i < (int)(sizeof(matches) / sizeof(matches[0])); i ++) | |
349 | if (matches[i].rm_so < 0) | |
350 | break; | |
351 | ||
352 | return (i); | |
353 | } | |
354 | else | |
355 | return (0); | |
356 | } | |
357 | ||
358 | ||
359 | /* | |
360 | * 'cgiFreeSearch()' - Free a compiled search context. | |
361 | */ | |
362 | ||
363 | void | |
364 | cgiFreeSearch(void *search) /* I - Search context */ | |
365 | { | |
366 | regfree((regex_t *)search); | |
367 | } | |
368 | ||
369 | ||
370 | /* | |
bc44d920 | 371 | * End of "$Id: search.c 6649 2007-07-11 21:46:42Z mike $". |
ef416fc2 | 372 | */ |